Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations Association Study of ABCA1 Polymorphisms in Singapore Populations 6.1 Introduction High density lipoprotein (HDL) levels are inversely correlated with the incidence of coronary artery disease (CAD; Wang and Briggs, 2004). A long-standing hypothesis to explain the atheroprotective effect of HDL is the process of reverse cholesterol transport (RCT; Glomset, 1968). In RCT, HDL or its apolipoproteins mediate the removal of excess free cholesterol from peripheral cells and, following a series of reactions in plasma, the cholesterol is delivered via either low density lipoprotein (LDL) or HDL to the liver for excretion into the bile. Until recently, little was understood about what primes the initial step of RCT, namely, the removal of cholesterol from peripheral cells. Moreover, evidence that definitively established the relationship between cholesterol efflux, HDL and atherosclerosis had been elusive. Tangier disease is a rare recessive disorder characterized by a near absence of plasma HDL-cholesterol (HDL-C) and a marked deposition of cholesterol esters in macrophage-rich tissues (Fredrickson et al., 1961). Biochemically, fibroblast cultures from Tangier disease patients are associated with a decrease in cholesterol (Walter et al., 1994) as well as phospholipid efflux (von Eckardstein et al., 1998). Although first described in 1961, the molecular basis for Tangier disease has only been recently solved as homozygosity in ABCA1 gene mutations (Bodzioch et al., 1999; Brooks-Wilson et al., 1999; Remaley et al., 1999; Rust et al., 1999). Certain patients with the milder and more common form of familial hypoalphalipoproteinemia, also known as familial HDL deficiency, are characterized by heterozygosity in ABCA1 gene mutations (Brooks-Wilson et al., 1999; Marcil et al., 1999). The recognition of ABCA1 gene defects as the molecular basis for these two forms of inherited HDL deficiencies has shed much light on the role of the ABCA1 transporter as the gatekeeper of cholesterol efflux. ABCA1 mediates in the 113 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations regulated and energy-dependent transfer of excess cholesterol and phospholipids from peripheral cells including macrophages in the arterial wall to acceptor molecules in the plasma such as ApoAI, thereby generating nascent HDL particles and igniting the process of RCT (Oram, 2003). The discovery of ABCA1 as a crucial mediator in the first step of RCT, together with the well established inverse relationship between HDL-C levels and atherosclerosis, provide the scientific notion that ABCA1 is the much sought after direct link between HDL and atherosclerosis. Direct evidence comes from observations that an impairment of cholesterol efflux is associated with increased risk of CAD and HDL-C seen in ABCA1 human heterozygotes (Clee et al., 2000; van Dam et al., 2003) as well as from transgenic mouse models (Singaraja et al., 2003). More generally, there is interest that the more common genetic variations in ABCA1 may explain phenotypic variability in HDL-C levels and/or CAD risk among individuals of the general population. To date, numerous association studies using singlenucleotide polymorphisms (SNPs) both in the coding and promoter regions of the ABCA1 gene have been reported in Caucasian, African-American and Japanese populations (Wang et al., 2000; Brousseau et al., 2001; Clee et al., 2001; Lutucuta et al., 2001; Zwarts et al., 2002; Cenarro et al., 2003; Evan et al., 2003; Harada et al., 2003; Kakko et al., 2003; Srinivasan et al., 2003; Yamakawa-Kobayashi et al., 2003; Shioji et al., 2004; Cohen et al., 2004; Frikke-Schmidt et al., 2004; Tregouet et al., 2004). A summary of the finding of these studies are compiled in Table 6.1. The results, however, have not been entirely consistent. To illustrate, for R219K, the most well-studied ABCA1 SNP to date, several studies show a protective association of the rarer K219 allele against CAD, which may or may not be accompanied by parallel increases in HDL-C and/or apolipoprotein AI (ApoAI) concentrations (Clee et al., 2001; Cenarro et al., 2003; Evans et al., 2003; Kakko et al., 2003;Tregouet et al., 2004; Yamakawa-Kobayashi et al., 2004). However, two recent large 114 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations sample studies (Cohen et al., 2004; Frikke-Schmidt et al., 2004) failed to replicate the protective association of the K219 allele. Similarly, studies involving the M883I locus have also been variable, ranging from no evidence of association to entirely contradictory effects (Table 6.1). Here, we have conducted an association study of seven ABCA1 SNPs spanning a region of ~145kb using population-based samples from Singapore. The seven SNPs examined in the association study were: -14C>T, 237indelG, R219K, V825I, M883I, IVS44+18T>C and 8995A>G (listed in order according to their chromosomal location). A schematic of the locations of the SNPs with respect to the genomic sequence and mRNA (for coding and UTR SNPs only) is shown in Figure 6.1. These SNPs were selected for various reasons. All are polymorphic in all three ethnic groups of Singapore with a minimum allele frequency of 5%. With the exception of 8995A>G, we were also restricted to evaluating SNPs which could be readily genotyped using simple restriction fragment length analysis without prior extensive assay development. Associations with CAD susceptibility and/or HDL-C had been previously reported for the coding SNPs R219K, V825I and M883I (Table 6.1), and it was therefore of interest to determine the replicability of those findings in the local context. We also studied -14C>T, as well as two novel SNPs discovered in the course of the study, IVS44+18T>C and 8995A>G. The promoter SNP 14C>T is less extensively investigated and nothing is known regarding the contributions of the intronic variant IVS44+18T>C and the 3’ untranslated region (UTR) variant 8995A>G to the traits of interest. In Singapore, Indians show the highest incidence of CAD followed by Malays and Chinese, and this risk difference parallels ethnic differences in HDL-C (Heng et al., 2000). This observation, in the context of a fairly homogeneous physical environment, may suggest the presence of a strong genetic component in determining CAD risk and HDL-C levels. Using a case-control approach, we compared the frequency distributions of ABCA1 115 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations SNPs between male CAD patients (cases) and their ethnic-matched healthy controls. In addition, a second analysis was carried out to explore the effect of these ABCA1 SNPs on quantitative variation in HDL-C, ApoAI and triglycerides (TG). The latter analysis was investigated only in control groups in order to exclude the effect of lipid-lowering treatment in cases. Both single- and multi-locus analyses were applied. The overall goal of the study was to establish the contributions of the selected ABCA1 SNPs to ethnic differences in CAD risk and lipid levels, especially HDL-C, among Singapore Chinese, Malay and Indian males. 116 Table 6.1 Published association studies of ABCA1 SNPs. ABCA1 SNP Variation Type Reference -564T>C Promoter Lutucuta et al. (2001) Study Population CAD patients Shioji et al. (2004) Tregouet et al. (2004) Ye et al. (2005) large Japanese cohorts Two UK-based populations UK white CAD patients -302C>T Promoter Shioji et al. (2004) large Japanese cohorts -278C>G Promoter -99C>G Promoter Zwarts et al. (2002) Shioji et al. (2004) Zwarts et al. (2002) Shioji et al. (2004) Zwarts et al. (2002) Shioji et al. (2004) Zwarts et al. (2002) Brousseau et al. (2001) Dutch CAD patients, cohort design large Japanese cohorts Dutch CAD patients, cohort design large Japanese cohorts Dutch CAD patients, cohort design large Japanese cohorts Dutch CAD patients, cohort design US case-control study Clee et al. (2001) Dutch CAD patients, cohort design Cenarro et al. (2003) German CAD patients Evans et al. (2003) German case-control study Harada et al. (2003) Japanese CAD patients Kakko et al. (2003) Srinivasan et al. (2003) Finnish population-based study US whites, population-based Cohen et al. (2004) Frikke-Schmidt et al. (2004) Shioji et al. (2004) Tregouet et al. (2004) Yamakawa-Kobayashi et al. (2004) Clee et al. (2001) Yamakawa-Kobayashi et al. (2004) Cohen et al. (2004) Frikke-Schmidt et al. (2004) Tregouet et al. (2004) Clee et al. (2001) Cohen et al. (2004) Frikke-Schmidt et al. (2004) Tregouet et al. (2004) Yamakawa-Kobayashi et al. (2004) Wang et al. (2000) Brousseau et al. (2001) US whites and blacks, general population Danish general population large Japanese cohorts Two UK-based populations Healthy Japanese young individuals Dutch CAD patients, cohort design Healthy Japanese young individuals US whites and blacks, general population Danish general population Two UK-based populations Dutch CAD patients, cohort design US whites and blacks, general population Danish general population Two UK-based populations Healthy Japanese young individuals Inuits, general population US case-control study Clee et al. (2001) Harada et al. (2003) Dutch CAD patients, cohort design Japanese CAD patients Kakko et al. (2003) Cohen et al. (2004) Finnish population-based study US whites and blacks, general population Frikke-Schmidt et al. (2004) Shioji et al. (2004) Tregouet et al. (2004) Yamakawa-Kobayashi et al. (2004) Clee et al. (2001) Danish general population large Japanese cohorts Two UK-based populations Healthy Japanese young individuals Dutch CAD patients, cohort design Cohen et al. (2004) Frikke-Schmidt et al. (2004) US whites and blacks, general population Danish general population Tregouet et al. (2004) Yamakawa-Kobayashi et al. (2004) Two UK-based populations Healthy Japanese young individuals -14C>T Promoter 237indelG R219K Promoter Coding V771M Coding V825I Coding M883I Coding R1587K Coding Association Trend towards lower HDL-C and ApoAI in heterozygotes Increased association with severity of atherosclerosis No association Marginally higher ApoAI for T allele Lower expression activity for T allele Increased CAD in TT homozygotes especially in smokers Association with higher HDL-C in general population but not in hypertension group Increased CAD events, no effect on lipids and genotypes Association with HDL-C in general population and hypertension group Fewer CAD events, no effect on lipids No association Increased atherosclerosis among T allele carriers No association Less severe atherosclerosis Higher frequency in cases No effect on lipids Decreased CAD severity, progression Decreased TG Increased cholesterol efflux and HDL-C in age-dependent manner Higher frequency in patients without premature CAD Interaction effect: Protection more among smokers than nonsmokers Lower K frequency in cases Decreased TG in cases who had not received lipid-lowering treatment, especially those with ApoE3/3 genotypes. No association with HDL-C and susceptibility to CAD Association with TG in K carriers, but dissipiated after adjusting for M883I, gender, BMI, smoking, hypertension and diabetes Higher HDL-C in women KK individuals No significant marginal effects on HDL Significant interaction of genotypes and age on HDL K carriers associated with higher HDL with age. No association with HDL-C No association with HDL-C No association with HDL-C Fewer myocardial infarction Increased HDL-C and ApoAI in KK individuals Decreased CAD severity in VM heterozygotes Higher HDL-C and ApoAI in M carriers. No association with HDL-C Higher HDL-C in VM heterozygotes in women No association with ApoAI Increased events in I carriers during trial Modest assocation seen in white men; direction of effect not specified Higher HDL-C in I heterozygotes in women No association with ApoAI No association with HDL-C, ApoAI, TG Higher HDL-C in MM homozygotes Higher frequency of M allele in cases compared to controls No association with CAD endpoints in cases. M carriers have a 9% reduction in TG in cases, but no such association in controls. Increased CAD severity in MM compared to II homozygotes Higher HDL-C in M carriers (omitting patients on lipid-lowering medication) No association with CAD susceptibility Modest association with higher HDL-C in women Modest association with higher HDL-C in black and white men homozygous for MM compared to II individuals. No association with HDL-C No association with HDL-C No association with ApoAI No association with HDL-C, ApoAI, TG Lower HDL-C in dose-dependent manner, independent of age, BMI, smoking and TG No effect on CAD No association with HDL-C Stepwise decrease in HDL-C and ApoAI over time, especially in women No difference in frequency between high and low HDL-C groups Lower ApoAI No association with HDL-C, ApoAI, TG 117 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations Genomic Organization -14C>T 237indelG R219K V825I M883I IVS44+18T>C 8995A>G 149 kb 24.3 kb 45.1 kb 32.8 kb 1.3 kb 33.1 kb 8.5 kb mRNA Organization 237indelG R219K V825I M883I 8995A>G 10.4 kb 0.73 kb 1.8 kb 0.2 kb 6.0 kb Figure 6.1 Relative locations of ABCA1 SNPs examined in association study. Spacing between SNPs is not drawn to scale. Promoter SNP (green), missense SNPs (red), UTR SNPs (blue), intronic SNP (black). 118 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations 6.2 Results 6.2.1 Characteristics of Study Subjects The demographic features and plasma lipid attributes of the association study subjects sampled from the Singapore Chinese, Malay and Indian populations are summarized in Table 6.2. Due to the small numbers of females recruited, only data from male subjects were analyzed. 6.2.1.1 Age Exploratory analyses indicated that age distributions for each sample followed a normal distribution, therefore parametric tests were used to examine age differences between groups. Cases were significantly older than controls in individual analyses for each ethnic population analysis (PC could also be in LD with the unobserved causative variant. When haplotypes were examined with variation in intermediate phenotypes, in Chinese controls, two extended haplotypes of 3g-R219-V825-I883-IVS44+18C-8995A and 3g-R219-I825-I883-IVS44+18T-8995A were found to be negatively and positively associated, respectively, with variation in HDL-C as well as ApoAI levels. In Indians, the haplotype 2g-K219-V825-I883-IVS44+18C-8995A was associated with raised TG levels, independent of HDL-C and ApoAI. On the other hand, sliding window analysis did not find any association. These findings of a significant extended haplotype association contrasts with the same analysis using the binary CAD trait in which haplotype effects were isolated when examining a smaller subset of SNPs in the Indian population. In the present study, age and smoking were non-genetic risk factors positively associated with CAD outcome, whereas BMI was negatively associated (Table 6.2). Although a positive association of age with a late-onset disease like CAD should not raise an alarm, the association is mostly likely due to a selection bias in the control samples. Smoking is a known modifiable CAD risk factor in Singapore populations (Lee et al., 2001). The statistically significant inverse association between BMI, which is a surrogate for obesity, and CAD risk seems contradictory (Lee et al., 2001) but the magnitude of the mean BMI difference was slight (Table 6.2) and also the association dissipated in Malays and Indians after factoring in age differences between cases and controls. We could speculate that there might be differences in diets between young and old subjects. Since cases were not recruited at baseline and many would have adopted a healthier lifestyle upon diagnosis, there is some basis to expect an inverse association between BMI and CAD risk. Contrary to expectations, mean levels of TC, LDL-C, ApoB were raised in controls compared to cases regardless of ethnicity, a finding which is likely due to the effect of lipid-lowering drug therapy received by cases (Table 6.2). Conversely, the 182 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations intermediate phenotypes of HDL-C, ApoAI as well as Lp(a) showed the expected differences between cases and controls (Table 6.2). One general observation of the present study is that none of the seven ABCA1 SNPs examined was consistently associated with either CAD status or the same quantitative lipid trait in all three ethnic groups. Uniform recruitment criteria were applied to all three ethnic groups and thus variability in phenotype definition is a slim possibility. This lack of a consistent genetic effect across all three ethnic groups in Singapore is not unknown (Tai et al., 2003; Qi et al., 2005) and may reflect differences in underlying genetic and non-genetic risk profiles between ethnicities. For instance, concentrations of LDL-C (and related ApoB), HDL-C (and related ApoAI), and Lp(a), and smoking habits vary among Chinese, Malays and Indians (Tables 6.2-6.3). Ethnic variations in established CAD risk factors in Singapore are well known (Lee et al., 2001). A second general observation is that we did not find any ABCA1 SNP that was concurrently associated with CAD and lipid phenotypes, especially HDL-C/ApoAI. For example, in Chinese, the 8995A>G locus was most significantly associated with CAD (Figure 6.5; see also Table 6.5) without a parallel effect on HDL-C, ApoAI or TG (Figure 6.10). This finding corroborates with many studies that had simultaneously examined ABCA1 SNP associations with clinical and lipid phenotypes (Brousseau et al., 2001; Clee et al., 2001, Harada et al., 2003; Zwarts et al., 2003; Tregouet et al., 2004; see also summary in Table 6.1). The lack of concordance could be related to the fact that atherosclerosis and related lipid phenotypes are complex traits regulated by overlapping as well as distinct sets of genetic factors. The list of candidate genes for atherosclerosis is quite extensive and includes those involved not only in lipid metabolism but also inflammation, vascular homeostasis and thrombosis (Lusis, 2000). Adding a further layer of complexity, non-genetic factors such as non-modifiable and lifestyle variables may also influence CAD and lipid phenotypes. In short, one might expect heterogeneity in 183 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations associations because of underlying differences in risk factor profiles. Some support for this comes from observations that certain ABCA1 SNP associations are restricted to selected subgroups such as women (Kakko et al., 2003; Frikke-Schmidt et al., 2004) or in nonsmokers (Cenarro et al., 2003; Tregouet et al., 2004). A criticism of the present study is that a limited number of non-genetic risk factors such as age, BMI and smoking was examined. Only data from males were analyzed and given the already small sample sizes, we did not perform stratified analysis based on smoking although the absence of a significant interaction effect between genetic covariates and smoking suggests that smoking does not modify the genetic association. Adequate collection of clinical characteristics will also render the full impact of gene-environment interactions to be evaluated (Sing et al., 2003). Given the multifactorial nature of complex traits like CAD and lipids, it will be crucial to obtain a more global picture of the contributions of the multiple genetic factors to be captured and to model their interactions (epistasis) in the analysis. Sample sizes, however, would have to be increased considerably in order to analyze epistatic as well as gene-environmental interactions (Wang et al., 2005). Nevertheless, the study does not detract from the primary goal of evaluating the contribution of common genetic variation in the ABCA1 gene to variability in CAD susceptibility and lipid concentrations in the general population. Recent mouse studies indicate that macrophage ABCA1 activity in the macrophage contributes minimally to plasma HDL levels but its inactivation in the macrophage does contribute significantly to risk of atherosclerosis (Haghpassand et al., 2001; Aiello et al., 2002). Instead, hepatic ABCA1 activity is implicated as a major source of HDL in plasma (Wellington et al., 2003; Basso et al, 2003). Thus a biologically feasible explanation for the absence of a concurrent association of CAD as well as plasma HDL-C may be related to the different roles of ABCA1 in macrophage and liver. 184 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations It may also be argued that specific endophenotypes such the small HDL subclasses, which are more intimately connected to the initial steps of RCT, may reveal the true contribution of ABCA1 SNPs to variation in HDL and CAD susceptibility by providing a stronger effect. Distinguishing and quantifying HDL subclasses, however, will require sophisticated analytical techniques such as nuclear magnetic resonance spectroscopy (Otvos, 2000). Routine biochemical analysis measures only HDL-C in plasma which represents a heterogeneous population of HDL particles. The contrasting associations of ABCA1 SNPs observed for the various ethnic groups in this study could be explained in part by differences in power. The power of a genetic association study depends on the sample size, allele frequencies, and the size of the risk to be detected. In this study, we estimated that the power to detect an association with the CAD trait ranged between ~63%-90% (average 76%) in Chinese, 35-62% (average 43%) in Malays and 12-71% (average 44%) in Indians, assuming a risk effect of OR of 1.5 based on a conventional 2x2 allele-based test (Table 6.4). Meta-analyses show that the most irrefutable disease susceptibility variants identified so far have modest allelic ORs on the order of 1.1-1.4 (Ioannidis, 2003; Lohmueller et al., 2003), and similarly, quantitative trait loci could contribute a similar magnitude of modest effect to a quantitative trait. Therefore, basing our power calculations on an effect size of OR of 1.5 seems unrealistic. The true power is likely to be further reduced because of a sample collection bias. On average, the healthy controls enrolled in this study were seven years younger than their corresponding ethnicity-matched cases (Table 6.2). Since CAD is a common late-onset disease and a proportion of the young controls may develop CAD later, a true allele frequency difference will be narrowed, thereby potentially masking the effect of an association. The impact might be especially severe for the Malay case-control dataset because Malay controls averaged 18 years younger than their CAD counterparts compared to a smaller mean age gap of ~11 years between cases and controls of Chinese 185 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations and Indian ethnicities. The retrospective nature of the case-control study meant that data were collected at a single time-point with no followup and the outcome of some controls could have been misclassified. Hence, the absence of an association cannot be safely dismissed due to an inherent sample collection bias. On the hand, it may be argued that the presence of an association might signal a stronger underlying effect. In this study, we have sought to limit the confounding effect of age through statistical adjustment. An oft-cited reason for the poor replicability of an association is because the associated SNP is not directly causal but rather in LD with the unobserved functional variant(s). In an association study, risk variants can be detected through direct or indirect assay of a marker that in LD with the risk variant (Collins et al., 1997). Recent data show that patterns of LD are highly variable between populations (Shifman et al., 2003 Crawford et al., 2004; Sawyer et al., 2005). These observations suggest that both false negative and false positive results in different populations must be interpreted carefully. To the extent that this true, we were unable to fully catalogue the patterns of sequence variation and LD in the ABCA1 gene in Chinese, Malays and Indians prior to embarking on the association study, and this will impact the efficacy of the association study. Our primary approach was to use an exon-centric approach for SNP discovery. In addition, we selected only seven markers with a minimum allele frequency of 5% in all three ethnic groups and which could be readily genotyped by available methods without requiring extensive assay development. Typing only seven markers in the ~149kb ABCA1 gene without prior assessment of their informational content with the rest of the common genetic variation in the gene will obviously give an inaccurate picture of LD. A detailed empirical assessment of LD in the ABCA1 gene would be instructive in two aspects. First, it would facilitate the selection of efficient tagging SNPs such that genotyping only a few carefully chosen SNPs in the region will provide sufficient information about the remaining common SNPs in that region (Carlson et al., 2004). Second, a high-resolution map of genetic markers can aid in 186 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations the understanding of how evolutionary processes such as natural selection, mutation, recombination and demography have moulded the present-day patterns of sequence variation. This population genetics approach has the potential to find a risk variant without actually having to find patients (Nachman et al., 2000; Gilad et al., 2002; Sabeti et al., 2002; Toomajian et al., 2003; Akey et al., 2004; Tang et al., 2004). A lack of replication in the ABCA1 SNP associations can also arise as a result of cryptic population structure or stratification. Population structure has the effect of not only creating false positives but can also mask or reverse the association (Deng, 2001). To what extent systematic differences in ancestry of cases and controls contribute to potential problems in association studies is hotly debated. Some authors argued that proper matching based on self-reported ancestry/race and geographical location should suffice (Ardlie et al., 2002b; Pankow et al., 2002; Risch et al., 2002; Wacholder et al., 2002) while others advocate that population structure cannot be safely ignored (Thomas and Witte, 2003; Ziv and Burchard, 2003; Marchini et al., 2004). Given that Singapore is a cosmopolitan, multi-ethnic society, population stratification is indeed a legitimate concern and should be addressed empirically and adjusted for statistically when present. This can be achieved by typing a panel of genetic markers unlinked to the locus of interest (Pritchard and Rosenberg, 1999) or ancestry-informative markers (Hoggart et al., 2003; Ziv and Burchard, 2003). Our study evaluated the role of several common genetic variations in the ABCA1 gene to inter-individual variability in CAD susceptibility and plasma lipids. Two recent well designed and large sample studies found that multiple rare coding ABCA1 variants with major phenotype effects also contributed substantially to variation in plasma HDL-C (Cohen et al., 2004; Frikke-Schmidt et al., 2004). Rare variants are also implicated in susceptibility to inflammatory bowel disease (Hugot et al., 2001; Ogura et al., 2001), meningococcal disease (Smirnova et al., 2003), and colorectal adenomas (Fearnhead et 187 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations al., 2004). Those results indicate that sequence-based approaches focusing on the direct identification and testing of rare genetic variants are needed in the association studies of complex traits as mapping by LD may be ineffective, although resequencing efforts still remain prohibitive for most laboratories. Furthermore, numerous SNPs in the ABCA1 gene are also known to exist (dbSNP; http://www.abca1-mutants.all.at), and it will be necessary to assess their contributions too. We are cautious not to emphasize the association results of the -14C>T locus. We noted a marked deviation from HWE of the -14C>T locus in several groups including controls, with an observed excess of CT heterozygous genotype and a deficit of the rare homozygous TT genotype (Table 6.4). The small P values of the HWE tests, the detection of the HWE deviation in multiple groups of reasonable sizes, and the general lack of HWE deviation at the other six loci eliminated the possibility of a statistical fluctuation or other reasons for HW disequilibrium such inbreeding, migration, population structure. The ABCA1 gene locus does not play a role in mate selection and consanguineous marriages are rare in Singapore. Initially, we had ruled out a genotyping error as both sequencing and RFLP gave completely concordant genotypes. There was also no evidence of mismatches in the PCR primers which would be expected to cause allele-specific amplification and a resulting deficit of heterozygotes, not an observed excess of heterozygotes. Furthermore, we were assured that the primers were amplifying a unique target. Nevertheless, technical reasons for the observed lack of HWE at -14C>T remain a possibility. Recently, a 5-bp insertion ACCCC polymorphism in the ABCA1 proximal promoter has been identified (Probst et al., 2004; Shioji et al., 2004). The sequencing as well as RFLP primers used for typing the -14C>T locus in this study flank the indel polymorphism but curiously, during the resequencing of the promoter, we did not notice overlapping double peaks at the point of insertion which is distinctive of indel-containing heterozygotes, although this might be dismissed due to the low prior probability of 188 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations detection in a small screening sample size (data not shown). With these additional observations, we now postulate that in the original -14C>T genotyping assay, the shorter deletion allele might be preferentially amplified, leading to a departure of HWE of genotypes at nearby SNPs. More extensive resequencing has since verified the presence of the indel polymorphism in local Chinese, Malays and Indians, and unusually, it is only detectable in the two homozygous forms without any heterozygotes (data not shown). Genotypes for the nearby promoter SNPs -99G>C, -278C>G, -302C>T, -407C>G and 564T>C (previously mentioned in Chapter 5) often failed the HWE test. SNP genotyping assays that amplify shorter segments excluding the indel polymorphism should show no such bias in amplification and therefore more suitable for genotyping ABCA1 promoter SNPs. A study conducted on UK Caucasians by Tregouet et al. (2004) found a similar frequency of the -14C allele (~64-65%) albeit without deviation from HWE. Interestingly, the -14C allele frequency varies markedly from the 86% observed in another Caucasianbased study (Zwarts et al., 2003), a discrepancy that could also be attributed to the fact that established CAD patients, who are unlikely to be representative of the general population, were used. In the presence of HWE departure, traditional 2x2 contingency table tests that look for an allele frequency difference between diseased and non-diseased group are not robust (Sasieni, 1999). Thus, we have opted to use genotype-based tests for association implemented under logistic regression, which not only removes the assumption of HWE but also has the additional advantage of adjusting for other covariates that can confound the ability to detect a genetic association. But the haplotype-based methods require the assumption of HWE in the EM inference of haplotype frequencies (Excoffier and Slatkin, 1995), and therefore as a precaution, we excluded the -14C>T locus when testing extended haplotypes for associations with phenotypes. 189 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations Another concern with the study relates to the amount of missing genotypic data. Missing data effectively reduce sample size and thus lower power to detect specific effects or increases in magnitude of the differential effects that can be detected. Starting with 8995A>G which was the first to be genotyped, we were able to analyze reasonable numbers of cases and controls, but wound up with rather small numbers for the subsequent SNPs, particularly R219K and V825I which were among the last SNPs to be typed (Table 6.4). Logistic concerns included degradation of older DNA samples as well as one-tube-one-genotype assay system which hampered throughout. Some multi-locus analytical methods (such as those analyzing CAD as a dependent trait and unphased genotypes as covariates, Tables 6.6-6.8; and the extended haplotype association tests, Tables 6.9 and 6.12) required completely genotyped individuals (i.e. listwise deletion). Imputation can be applied to fill in gaps in data but this is best suited to situations in which there are relatively few missing values (Badzioch et al., 2003). To deal with this reduction in sample size, we adopted certain partial solutions: sliding window haplotype analysis consisting of smaller subsets of SNPs, and for the multi-locus association analysis with lipids using unphased SNPs, we selected genetic covariates by a purposeful selection strategy in which the most significant univariable SNP was used as a covariate. We found the locus-based multi-locus analysis, which we applied for the association test with CAD trait, generally reaffirms the result of the single-locus analysis in that the genetic main effects in the multivariable model corresponded to the top associated SNPs in single-locus tests. This finding is quite encouraging despite a much reduced dataset in the multi-locus analysis, and suggests evidence of allelic heterogeneity in contributing to CAD risk. Throughout, we have adopted a statistical analytical approach that is parsimonious, meaning that interaction effects were not fitted into the model unless individual marginal (main) effects were identified. Thus the analysis does not consider a situation in which the genetic covariate has no overall effect on its own despite having strong effects within 190 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations certain strata of the samples defined by other genetic or non-genetic covariates. Although taking into account possible statistical interactions could potentially increase the power to detect a novel causal variant, the power gain is likely to be attenuated by the need to protect against false positives due to subgroup analysis (Wang et al., 2005). Also, the presence of statistical interaction may not easily translate into an etiological meaning (Clayton and McKeigue, 2001). In summary, we have tested several ABCA1 SNPs for association with CAD and intermediate phenotypes of HDL-C, ApoAI and TG, in the three predominant ethnic populations of Singapore. Results reveal the presence of ethnic-specific associations. These observations extend both to single- as well as multi-locus analyses. While the findings may suggest a genuine allelic heterogeneity in which different ABCA1 genetic risk variants predominate in different populations, other reasons cannot be excluded including heterogeneity in power and genetic as well as non-genetic backgrounds, presence of stratification within samples, and perhaps even chance findings due to repeated testing. Therefore, recommendations for future studies would be to adopt an optimal epidemiological study design such as a large and well-defined collection of cases and controls with replication in an independent sample and/or functional verification of the association, perform a detailed assessment of the pattern of sequence variation in the gene, as well as explore empirical issues such as the extent of population structure among local ethnic populations. Such measures should enable a thorough assessment of the contributions of ABCA1 genetic variations to phenotypic variability in CAD susceptibility and HDL-C levels in the general population. Given the essential role of ABCA1 in RCT, it is also anticipated that there will be more ABCA1 SNP association studies conducted in the future, and the accumulating data could potentially be combined in a meta-analysis, which offers one means of achieving sufficient sample sizes for detecting associations with small effects. With accessibility to modern genotyping platforms, the scope of the analysis could 191 Chapter Association Study of ABCA1 Polymorphisms in Singapore Populations also be expanded to other candidate genes lying in the causal pathways of atherosclerosis and HDL level regulation. These approaches should greatly enhance our understanding about the basis for the contrasting cardiovascular risks among Singapore Chinese, Malays and Indians. 192 [...]... -6. 235 1.959 0.002 -1.797 2.2 16 0.418 0.190 0. 063 0.003 -0.142 0.183 0.4 36 1.718 3.4 06 0 .61 4 Lp(a) β SE P 0.028 0.1 06 0.793 0.340 0.0 86 0.05, Table 6. 4) These -14C allele frequencies for the three diverse Singapore populations. .. 71.5% 98.0% 1 26 (71 .6% ) 70 (61 .9%) 44 (25.0%) 40 (35.4%) 6 (3.4%) 3 (2.7%) 1 76 113 0.841 0.7 96 n.s n.s 35.0% 72.7% 115 (49.8%) 92 (57.5%) 93 (40.3%) 51 (31.9%) 23 (10.0%) 17 (10 .6% ) 231 160 0 .69 9 0.734 n.s n.s 68 .0% 98.0% Chinese Control Chinese CAD 53 (39.0%) 67 (44.1%) 63 ( 46. 3%) 63 (41.4%) 20 (14.7%) 22 (14.5%) 1 36 152 0 .62 1 0 .64 8 n.s n.s 62 .8% 96. 7% 48 (40.0%) 25 (47.2%) 55 (45.8%) 21 (39 .6% ) 17 (14.2%)... OR of 1.5 The number of Indian cases and controls genotyped had the weakest power (40.5%) to detect a modest crude allelic OR of 1.5, but 135 Chapter 6 Association Study of ABCA1 Polymorphisms in Singapore Populations was otherwise reasonably powered ( 86. 3%) for a larger OR of two, assuming a significance level of 5% Association test results for CAD at the M883I locus are presented in Table 6. 5 In. .. out of the seven ABCA1 loci (237indelG, R219K, V825I, IVS44+18T>C and 8995A>G) Single-locus association analyses for these trimmed datasets were not conducted due to an inherent reduction in power associated with small samples 145 Chapter 6 Association Study of ABCA1 Polymorphisms in Singapore Populations A Chinese CAD 1 2 3 1 2 3 4 5 6 7 Malay CAD 1 2 3 1 2 3 4 5 6 7 Indian CAD 1 2 3 1 2 3 4 5 6 7... three single-locus association test results (age-adjusted PC IVS44+18T>C is a T to C transition polymorphism located within intron 44 of the ABCA1 gene, and was discovered during the DHPLC survey (Figure 5.13) The polymorphism is located 18 bases from the 5’ splice site of exon 44 and is not known to disrupt consensus . Chapter 6 Association Study of ABCA1 Polymorphisms in Singapore Populations 113 6 Association Study of ABCA1 Polymorphisms in Singapore Populations 6. 1 Introduction High density lipoprotein (HDL). error. Chapter 6 Association Study of ABCA1 Polymorphisms in Singapore Populations 129 6. 2.2 Single-Locus Associations with CAD In the following subsections, the associations of each SNP with. allele was present at frequencies of 65 % in both Chinese and Indians, and 68 % Chapter 6 Association Study of ABCA1 Polymorphisms in Singapore Populations 130 in Malays, with no significant differences